Sidekick - Local AI Assistant

Sidekick brings the power of local AI directly into your browser. Chat with your own locally-running language models without sending any data to the cloud. Your conversations stay on your machine, giving you complete privacy and control. v2.1 updates: 1. Vision/Image Support - Paste images (Ctrl/Cmd+V) directly into the chat - Works with multimodal models (gemma3, llava, etc.) - Image preview with remove button before sending - Auto-switches to /v1/chat/completions endpoint when image attached 2. System Dark Mode - Auto-detects macOS/Windows light/dark preference - Dark theme for sidebar chat and settings page - Native dark scrollbars and form controls 3. Minor Improvements - Better error messages for connection failures - Updated input placeholder: "Use /commands or Ctrl+V to paste images" ━━━━━━━━━━━━━━━━━ 🤖 WHAT IS SIDEKICK? Sidekick is a browser sidebar that connects to your local LLM server. Think of it as having ChatGPT-like capabilities, but running entirely on your own hardware. No API keys needed. No subscription fees. No data leaving your computer. Whether you're summarizing articles, drafting professional emails, extracting action items from meeting notes, or just having a conversation - Sidekick makes it easy to leverage AI while browsing. ━━━━━━━━━━━━━━━━ ✨ KEY FEATURES 【Sidebar Interface】 Sidekick opens as a convenient sidebar panel, so you can chat with AI while viewing any webpage. No popups or new tabs - everything stays in one place. 【Text Selection Integration】 Select any text on a webpage and Sidekick automatically detects it. Perfect for asking questions about content, getting summaries, or transforming text. 【Quick Actions】 Use the dropdown menu or slash commands to quickly: • /summarize - Get concise summaries of long content • /explain - Break down complex topics in simple terms • /professional - Rewrite text in a professional tone • /actions - Extract action items and to-dos • /twitter - Convert content into a tweet thread 【Slash Command Autocomplete】 Type "/" to see all available commands with descriptions. Use arrow keys to navigate and Tab to complete - just like a modern code editor. 【Streaming Responses】 See responses as they're generated in real-time, just like ChatGPT. No waiting for the complete response before seeing any output. 【Conversation History】 Sidekick remembers your conversation context, allowing for natural back-and-forth dialogue. Clear the chat anytime to start fresh. 【Stop Generation】 Changed your mind? Click the stop button to halt response generation at any time. 【Markdown Support】 Responses are beautifully formatted with full markdown support - headings, bold, italics, code blocks, lists, and more. ━━━━━━━━━━━━━━━━ 🔒 PRIVACY FIRST Unlike cloud-based AI services, Sidekick keeps everything local: • Your prompts never leave your computer • Your conversations are not stored on any server • No account required • No API keys to manage • No usage limits or quotas • No monthly fees This makes Sidekick perfect for: • Working with sensitive or confidential information • Corporate environments with strict data policies • Users who value privacy and data ownership • Anyone who wants AI assistance without the cloud ━━━━━━━━━━━━━━━━ 🖥️ SUPPORTED LLM SERVERS Sidekick works with any server that supports the OpenAI-compatible API format (/v1/chat/completions endpoint): 【Ollama】 The most popular way to run LLMs locally. Free and open source. • Default address: http://localhost:11434 • Install from: https://ollama.ai • Run: ollama pull gemma3:1b 【LM Studio】 User-friendly desktop app for running local LLMs with a nice GUI. • Default address: http://localhost:1234 • Enable "Local Server" in settings • Download from: https://lmstudio.ai 【llama.cpp】 Lightweight, high-performance LLM inference. • Run with the --api flag • Great for advanced users 【vLLM】 High-throughput LLM serving. • Default address: http://localhost:8000 • Excellent for multi-user setups 【Text Generation WebUI (Oobabooga)】 Feature-rich web interface for LLMs. • Enable the API extension 【AnythingLLM】 All-in-one AI application. • Check server settings for API address 【Any OpenAI-Compatible Server】 If it supports /v1/chat/completions, it works with Sidekick! ━━━━━━━━━━━━━━━━ 📋 QUICK START GUIDE 1. Install a Local LLM Server The easiest option is Ollama: • Download from https://ollama.ai • Install and run it • Open terminal and run: ollama pull gemma3:1b 2. Configure Sidekick • Click the Sidekick icon in your browser toolbar • Click the settings gear icon • Enter your server address (e.g., http://localhost:11434) • Enter your model name (e.g., gemma3:1b or google/gemma-3-1b) • Click "Test Connection" to verify 3. St…